How to Fish patches and reviews: does the score follow updates
Contents
Six patch notes landed in the first eight days after release. The obvious question: can you see them in the daily score? We merged two Steam tables — the day-by-day review histogram and the patch chronicle — into one.
A warning first, and it matters more than the table itself: this is a coincidence in time, not a proven effect. We explain why below.
The table
| day | positive | reviews that day | patch that day |
|---|---|---|---|
| 20.08 | 96.3% | 767 | release |
| 21.08 | 94.6% | 1,854 | 1.0.4 |
| 22.08 | 93.7% | 3,593 | 1.0.5 |
| 23.08 | 93.6% | 4,535 | 1.0.6 |
| 24.08 | 94.4% | 4,455 | 1.0.8 and 1.0.9 |
| 25.08 | 93.8% | 4,007 | — |
| 26.08 | 93.8% | 3,999 | — |
| 27.08 | 95.2% | 4,351 | 1.0.10 |
| 28.08 | 95.4% | 3,436 | — |
| 29.08 | 95.1% | 3,340 | — |
| 30.08 | 95.5% | 3,237 | — |
31 August is deliberately absent: the day was not over, and its figure is incomplete.
What is visible at a glance
The minimum is 23 August, 93.6%. That is the day patch 1.0.6 shipped. The post-release maximum is 30 August, 95.5%. That is the third consecutive day without any patch. The whole amplitude across 11 days is 1.9 percentage points.
Averaging by group:
- days a patch shipped (21, 22, 23, 24 and 27 August) — 94.30%;
- days without a patch (25, 26, 28, 29 and 30 August) — 94.72%.
So on patch days the score was on average 0.42 points lower.
Why you cannot conclude “patches hurt”
The two groups of days are separated not by patches but by time. All the patch days are early; all the patch-free days (except 25-26 August) are late. And the score rises over time on its own anyway: the first wave of buyers is always more impulsive than the ones that follow.
You can see it in the third column: the peak by review volume is 23 August (4,535), and that same day produced the worst share. Then the flow weakens — 3,237 reviews on 30 August — while the positive share climbs.
In other words time and patches changed together, and this data cannot separate them. Proving a patch effect would require a comparison group — players who did not receive that patch. No such group exists.
The one coincidence worth noting
25 and 26 August were two consecutive patch-free days with an identical 93.8%. On 27 August 1.0.10 shipped, and the same day the share jumped to 95.2% — the biggest single-day rise in the sample, +1.4 points.
Even here we claim no cause: reviews that day were also written by people who had not yet seen the patch, and Steam does not show a “before / after update” breakdown.
What is actually useful here
One observation does hold regardless of cause: across 11 days the daily positive share never once fell below 93.6%. Neither launch problems nor patches produced a day you could call a failure. And the all-time score is 94.9%, labelled Very Positive.
Frequently asked questions
Did patches improve the How to Fish score?
Only a coincidence in time was measured. Patch days averaged 94.30% positive, patch-free days 94.72%. The data proves no cause.
Which day scored worst?
23 August — 93.6% positive across 4,535 reviews. That is both the day patch 1.0.6 shipped and the peak of the review flow.
Did the score ever drop below 93.6%?
No. Across 11 days no single day produced a worse result.
Why can you not say patches raised the score?
Because time and patches changed together. Proving an effect needs a group of players without the patch, and no such group exists.